Alcohol withdrawal syndrome (AWS) carries a significant risk of morbidity and mortality yet has not been well- studied epidemiologically or therapeutically. In the hospital setting, up to 40% of patients have alcohol-use disorders that put them at risk for developing AWS, yet little is known about the national impact of AWS in the hospitalized setting. Few studies have examined the prevalence of AWS-related inpatient admissions nor trends in utilization over time. In terms of treatment, benzodiazepine administration remains the mainstay of current AWS care. However, inappropriate utilization and application of this treatment approach is easy in the inpatient setting, prompting evaluation of alternative treatment options including anticonvulsants such as phenobarbital. The aims of this project are to establish the prevalence and trends in inpatient AWS-related admissions, while also identifying patient- and hospital-level determinants of outcome and utilization. We will investigate these questions using 2000 - 2012 data from the National Inpatient Sample (NIS), an all-payer, nationally-representative dataset with approximately 7 million annual hospital discharges. Building upon these findings, we will then investigate the effectiveness of phenobarbital in treating AWS in the intensive care unit (ICU) in a quasi-experimental intervention conducted in the Beth Israel Deaconess Medical Center medical ICUs. We hypothesize that while AWS-related hospitalizations have remained stable between 2000 - 2012, the burden of comorbidities has increased and significant variability in resource utilization costs, and outcomes exists. We further hypothesize that a phenobarbital-based treatment protocol will be associated with decreased time to resolution of AWS, as well as decreased ICU and hospital length-of-stay, in critically ill patients with AWS. As part of this post-doctoral research training program, the principal investigator will complete a Master of Public Health with a concentration in Clinical Effectiveness to obtain the necessary training in epidemiology, biostatistics, and research methodology to complete this work, which will provide crucial experience in nationally-representative health services research and locally-based effectiveness evaluation along with content education in acute alcohol-related pathologies. This research project will be performed under the supervision and mentorship of an investigator well established in the field of the health effects of alcohol intake, with support from an expert in large database analysis and other experts in critical care.